GPT Researcher MCP Server

Content

GPT Researcher MCP Server connects Claude and other MCP clients to GPT Researcher through a dedicated MCP server repository. It exposes tools for deep web research, faster search, report writing, source retrieval, context retrieval, and reusable research resources.

Use it when Claude needs a supervised research workflow that can gather multiple sources, return citations, and produce a draft report instead of only returning a short web-search result list.

Source Review

• https://github.com/assafelovic/gptr-mcp
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/README.md
• https://github.com/assafelovic/gpt-researcher
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/LICENSE
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/server.py
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/utils.py
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/requirements.txt
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/.env.example
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/Dockerfile
• https://raw.githubusercontent.com/assafelovic/gptr-mcp/master/docker-compose.yml

These sources were reviewed on 2026-06-06. Prefer the dedicated MCP repository, README, linked GPT Researcher upstream, license, server source, utilities, dependency list, environment example, Dockerfile, and compose file for current setup and behavior details.

Features

• Run deep_research for multi-source web research with source URLs and context.
• Run quick_search for faster search-result snippets.
• Generate reports from a previous research session with write_report.
• Retrieve research sources and full research context by research ID.
• Expose a research://{topic} resource for reusable research context.
• Use a research_query prompt to shape a research task and report goal.
• Run over local stdio for desktop MCP clients.
• Use Docker/SSE or Streamable HTTP transport for containerized or web-oriented integrations.

Installation

Clone the dedicated MCP server repository and install its Python dependencies:

git clone https://github.com/assafelovic/gptr-mcp.git
cd gptr-mcp
pip install -r requirements.txt

Set provider credentials in the MCP client configuration. For a local stdio client, point to server.py:

{
  "mcpServers": {
    "gptr-mcp": {
      "command": "python",
      "args": ["REPLACE_WITH_GPTR_MCP/server.py"],
      "env": {
        "OPENAI_API_KEY": "REPLACE_WITH_OPENAI_API_KEY",
        "TAVILY_API_KEY": "REPLACE_WITH_TAVILY_API_KEY"
      }
    }
  }
}

The server can also run directly:

python server.py

Docker mode auto-detects container execution and switches to SSE transport on port 8000.

Use Cases

• Ask Claude to research a current topic, company, market, library, or technical question with source-backed context.
• Generate a report from previously gathered research context.
• Retrieve the source list or full research context behind an answer.
• Compare quick search output against deeper research before spending more API budget.
• Connect a Dockerized MCP deployment to n8n or another client that expects SSE endpoints.

Safety and Privacy

GPT Researcher MCP Server is a networked research tool. Treat every query as data that may be sent to model providers, search retrievers, and fetched web sources. Keep prompts free of secrets and review provider retention rules before using it for customer, legal, medical, financial, or internal strategy research.

For local use, protect MCP client config files that contain API keys. For Docker, n8n, SSE, or Streamable HTTP deployments, bind only to trusted networks, isolate containers, and avoid exposing research sessions or /messages endpoints without an authentication layer.